95598 customer appeal processing method and device and storage medium

Abstract

The invention discloses a 95598 customer appeal processing method and device and a storage medium. The processing method comprises the following steps: S101, data word segmentation: preprocessing voice translation text data of a customer, performing sentence segmentation on the data, and performing word segmentation on words of each sentence to obtain result data after word segmentation; s102, performing data vectorization, and generating word vectors by result data after word segmentation; s103, word vector processing: inputting the generated word vector into a customer appeal recognition algorithm model, extracting the core semantics of the text through a neural network structure, capturing the semantics of different words / words in the whole sentence meaning in the whole sentence meaning, and recognizing the appeal classification of the voice translation text data of the customer according to the semantics; and S104, outputting a result, and sending the result of the appeal classification to a corresponding processing end. According to the method, extraction of text semantics and generation of features can be realized under relatively low computing resources, and then different text semantic features are compared, so that the purpose of judging text contents is achieved.

Description

technical field [0001] The invention belongs to the technical field of computer communication, and relates to a 95598 client request processing method, device and storage medium based on multi-model fusion. Background technique [0002] With the continuous development of the social economy and the improvement of living standards, the consumption awareness of power customers has become stronger and stronger. The traditional power supply services have been unable to meet the daily needs of customers, and power customers are increasingly demanding diversified and differentiated services. At the same time, due to the improvement of marketization in other industries and the increasing improvement of the overall service level of the society, power customers have more and more types of demand for power supply services, and their expectations are also getting higher and higher. [0003] As the unified hotline service window of State Grid Corporation of China, the 95598 hotline has a...

AI Technical Summary

Problems solved by technology

However, the current traditional statistical-based technical solutions cannot achieve accurate semantic classification, and thus cannot realize the development of precise services

[0007] Based on actual engineering experience, a layer of BiLSTM network is used to process the corpus. Since some phone conversation sentences are very long, all the features cannot be effectively retained, so the classification and recognition effect is not good. Self-Attention network is used for feature extraction of long sentences , has better performance, and the current popular method is to use a multi-layer Self-Attention combination for semantic extraction. Google's Bert is a 12-layer Self-Attention combination, but the disadvantage of this network structure is that it is expensive for computing resources. It takes up a lot of space and requires professional, multiple GPU servers to get the ideal processing speed

Images